Search

Pattern matching is a fundamental problem in theoretical computer science. The algorithms for pattern matching and the study on the statistics of pattern matching have found enormous applications in practical fields. In this paper, we revisit the Markov embedding approach for studying pattern matching in repeated experiments. For any pattern of length $m$ over alphabet $\Sigma $ , we show that the mean and variance of the waiting time of the pattern in iid experiments can be computed in $O(m)$ time based on Markov embedding technique, improving over the $O(|\Sigma | \cdot m)$ and $O(m^{2})$ naïve bounds. Our method extends to computing the $k$ -th moment of the waiting time, and it extends to computing other related statistics about pattern matching in repeated experiments, and it also extends to the case of Markov dependent experiments.

In order to compensate for any failure on the use of point spread function (blur kernel) estimation and image estimation priors, we propose a novel regularization priors scheme with adapting the parameter for image restoration involving adaptive optics (AO) images. Our scheme uses a maximum a posteriori estimation with Gaussian statistics on the image and point spread function (blur kernel). An efficient regularization prior method associated with alternating minimization method is described to obtain the optimal solution recursively. Our method is applied to synthetic and real adaptive optics images. After applying our restoration method, satisfying results are obtained. Experimental results demonstrate that our proposed model and method performs better for restoring images in terms of both subjective results and objective assessments than the current state-of-the-art restoring methods. In addition, our proposed method can be a new way to promote their performances for AO image restoration.

An acceptance sampling plans are statistical tools in quality control which often used for lot inspection in several areas such as industry, engineering and business. It can be applied for preserving the quality of products in industry process and preserving the producer's risk and consumer's risk in the production process of manufactures. The objective of this study is to utilize the Empirical Bayes approach based on squared error loss and precautionary loss functions for parameter estimation in sequential sampling plans. The parameters are estimated using Lindley's approximation technique, and hyper-parameters can be obtained via Gibbs sampling technique. Data are normally distributed under an unknown mean and variance. The proposed plans are compared with traditional approaches including a single sampling plan and sequential sampling plan. The probability of acceptance (Pa) and average sample number (ASN) are used as criterion for comparison. Results show that the proposed plans yielded the smaller ASN and higher Pa than both classical methods.

This study focuses on modeling, prediction, and analysis of confirmed, recovered, and death cases of COVID-19 by using Fractional Calculus in comparison with other models for eight countries including China, France, Italy, Spain, Turkey, the UK, and the US. First, the dataset is modeled using our previously proposed approach Deep Assessment Methodology, next, one step prediction of the future is made using two methods: Deep Assessment Methodology and Long Short-Term Memory. Later, a Gaussian prediction model is proposed to predict the short-term (30 Days) future of the pandemic, and prediction performance is evaluated. The proposed Gaussian model is compared to a time-dependent susceptible-infected-recovered (SIR) model. Lastly, an analysis of understanding the effect of history is made on memory vectors using wavelet-based denoising and correlation coefficients. Results prove that Deep Assessment Methodology successfully models the dataset with 0.6671%, 0.6957%, and 0.5756% average errors for confirmed, recovered, and death cases, respectively. We found that using the proposed Gaussian approach underestimates the trend of the pandemic and the fastest increase is observed in the US while the slowest is observed in China and Spain. Analysis of the past showed that, for all countries except Turkey, the current time instant is mainly dependent on the past two weeks where countries like Germany, Italy, and the UK have a shorter average incubation period when compared to the US and France.

The tree height and crown diameter are important measurement attributes in forest resource survey and management. Hence, we propose a passive measurement method of tree height and crown diameter based on monocular camera of a smartphone. First, we use an feature-adaptive Mean-Shift algorithm to segment the image and extract tree’s contour. Furthermore, an adaptive feature coordinate system is established to help study the conversion relationship of the coordinate systems. It has been proved that for the image points with the same abscissa pixels, their ordinate pixels have a linear relationship with its actual imaging angles. A depth extraction model is built according to this principle. Then, we obtain the rotation and translation matrix and established tree height and crown diameter models according to the mapping transformation relationship of coordinates. Experimental results reveal significant correlation between calculated and truth values. The RMSE is 0.267 m (rRMS=2.482%) for tree height and 0.209 m (rRMS=5.631%) for crown diameter. The relative errors of tree heights are less than 5.76% (MRE=2.159%); for crown diameter, the relative errors are less than 9.73% (MRE=4.95%). Overall, the accuracy of this method falls within the requirements of the continuous inventory of Chinese national forest resources.

The steady-state visual evoked potential (SSVEP) visual acuity is usually defined by extrapolating a straight line regressed through significant SSVEP amplitudes plotted versus spatial frequencies to 0 μV or a noise level floor, or the finest spatial frequency evoking a significant SSVEP. This study aimed to compare the performance of the commonly used threshold determination criteria of the extrapolation technique and the finest spatial frequency technique. Visual acuity was measured both by the Freiburg Visual Acuity Test (FrACT) and SSVEP with vertical sinusoidal reversal gratings in ten adults. The extrapolation technique including three methods of linear extrapolation to zero (C1), linear extrapolation to noise level baseline (C2) and linear extrapolation to zero versus log spatial frequency (C3), and the finest spatial frequency technique with significance determination by canonical correlation analysis (CCA) and “OR”operation (C4) were used to determine the SSVEP visual acuity. Bland-Altman method found a pretty good agreement between the SSVEP and FrACT acuity obtained by all the four threshold estimation criteria. One-way repeated-measures ANOVA and Bonferroni post-hoc analysis found that there was no significant difference among visual acuities measured by FrACT and all the four criteria, except for the visual acuity estimated by C1 slightly higher than that of C2, demonstrating that these visual acuity estimating methods had a similar performance in evaluating the visual function. The correlation and agreement between subjective FrACT acuity and objective SSVEP acuity measured by four criteria respectively were all pretty good, demonstrating that all of these four threshold estimation criteria had a good performance in SSVEP visual acuity assessment.

Recently, enthusiastic studies have devoted to texture synthesis using deep neural networks, because these networks excel at handling complex patterns in images. In these models, second-order statistics, such as Gram matrix, are used to describe textures. Although these models have achieved promising results, the structure of their parametric space is still unclear. Consequently, it is difficult to use them to mix textures. This paper addresses the texture mixing problem by using a Gaussian scheme to interpolate deep statistics computed from deep neural networks. More precisely, we first reveal that the statistics used in existing deep models can be unified using a stationary Gaussian scheme. We then present a novel algorithm to mix these statistics by interpolating between Gaussian models using optimal transport. We further apply our scheme to Neural Style Transfer, where we can create mixed styles. The experiments demonstrate that our method outperforms a number of baselines. Because all the computations are implemented in closed forms, our mixing algorithm adds only negligible time to the original texture synthesis procedure.

Aiming at the problem of cement grinding specific surface area (SSA) detection, a cement SSA model based on XGBoost is proposed. First, based on actual production data of a cement plant, original data missing values and outliers are detected and processed. Secondly, using Pearson Correlation Coefficient (PCC) and Mutual Information (MI) method for feature selection, the multicollinearity between auxiliary variables is detected, and auxiliary variables with a high correlation to the cement SSA are screened out to achieve dimensionality reduction of input variables. Finally, the prediction model of cement SSA based on XGBoost is established. The results show that, compared with the GBM and RF, XGBoost has obtained a higher square of correlation coefficient (R2) and smaller root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE).

In this paper, our goal is to model the aggregate mobility of individuals in a city by analyzing cellular network connections, and then leverage the designed mobility model to model and predict the number of COVID-19 infections in future. We analyze cellular network connections from 973 antennas for all users in the city of Rio de Janeiro from April 5, 2020 to July 2, 2020. We design a Markovian model that captures the mobility across municipalities. We then combine the transition probabilities of the Markov chain with the number of COVID-19 cases in a municipality during a particular week in the design of our mobility-aware COVID-19 case prediction models to predict the number of cases for the following week. Our experiments demonstrate that our mobility-aware models significantly out-perform a baseline mobility-agnostic linear regression model in terms of metrics such as Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE).

In this paper, we introduce the logit Pearson type III distribution and we utilize it, for first time in the literature, to investigate the statistical behavior of wireless power transfer, assuming that the harvested energy follows a well-established nonlinear energy harvesting (EH) model based on the logistic function. Specifically, we present closed-form expressions for the statistical properties of the introduced logit Pearson type III distribution, e.g., the cumulative distribution function (CDF), the probability density function and the moments and we utilize this distribution to define the logit gamma distribution. Furthermore, taking into account that the logit Pearson type III distribution is closely related to the considered nonlinear EH model the statistical properties of the distribution of the harvested power are derived. These expressions are of high practical value, since useful insights for the EH system can be extracted through the evaluation of the CDF, as well as the average harvested power and the variance.

Making accurate and effective weather forecasts has been one of the important issues in human history. Its importance is further increased by its direct impact on economic areas such as agriculture and tourism, and its effects on saving people’s lives by predicting high-level air disasters. Especially today, when global warming has increased its impact on weather events, sudden weather changes and subsequent disasters have made it very important to develop accurate weather forecast methods and models. In this study, a deep learning based prediction model has been developed for weather forecast. The developed LSTM (Long Short-Term Memory) based deep learning model has been compared implementally with GRU (Gated Recurrent Unit), one of the recurrent neural network models, and traditional machine learning algorithms such as RF (Random Forest) and SVR (Support Vector Regression). For each model and algorithm, the experimental results obtained according to MSE (Mean Squared Error), RMSE (Root Mean Squared Error), MAE (Mean Absolute Error), Pearsons correlation coefficient and R2 metrics have been compared. Experimental results show that the developed LSTM-based deep learning model gives more successful results than other compared models and algorithms.

As data continues to grow rapidly at an exponential rate, the importance of data processing systems to extract information is also increasing continuously. Recommender Systems are information processing software that seek to aid decision making by providing relevant suggestions from a given pool of items. In this domain, earlier studies have presented a comparison of similarity measures mostly using Root Mean Squared Error and Mean Absolute Error and to the best of our knowledge, they lacked analysis based on the computation time. Therefore, in this paper, we present a comparison of three similarity measures, namely Pearson Correlation Similarity, Cosine Vector Similarity, and Mean Square Difference on two different datasets known as MovieLens 100k and MovieLens 1M and compared the results using three accuracy metrics namely, Root Mean Square Error, Mean Absolute Error, and Fraction of Concordant Pairs. Our results show that in terms of accuracy, Pearson Correlation Similarity is the best performing similarity measure. Moreover, we also consider the time of simulations for each measure to add another dimension to performance evaluation. As a result, Mean Squared Difference was found to be the fastest among all the three similarity measures

On June 29 2021, the World Health Organization (WHO) reported around 45,951 confirmed cases and 817 deaths of COVID-19 in India, and 64,903 confirmed cases and 1,839 deaths in Brazil. This virus has been determined as a global pandemic by WHO. Accurate forecast of COVID-19 cases has become a crucial task in the decision-making of hospital managers to optimally manage the available resources and staff. In this study, the Gaussian process regression (GPR) model tuned by Bayesian optimization (BO) was used to forecast the recovered and confirmed COVID-19 cases in two highly impacted countries, India and Brazil. Specifically, the BO algorithm is employed to find the optimal hyperparameters of the GPR model to improve the forecasting quality. We compared the performance of the Optimized GPR with 14 models, including Support vector regression with different kernels, GPR with different kernels, Boosted trees, and Bagged trees. We also applied the BO to the other investigated predictors to maximize their forecasting accuracy. Three statistical criteria are used for the comparison. The daily records of confirmed and recovered cases from Brazil and India are adopted in this study. Results reveal the high performance of the GPR models compared to the other models.

The weights have a very large influence on the parameter estimation of the geographically weighted regression (GWR). The weights show the relationship between observations or locations in the model. Types of weights that are often used in GWR are Gaussian kernels. This weighting can also be arranged into two forms. There are the fixed Gaussian kernel and the adaptive Gaussian kernel. Fixed is used when each location has the same bandwidth value. Adaptive is used when each location has a different bandwidth value. This study compares the results of the estimated parameters of the GWR model using the two weights. The modelling is carried out using data on maternal mortality in Central Sulawesi Province, consisting of 13 regencies/cities. The data used is secondary data. The result is that the adaptive Gaussian kernel weighting gives better results on the GWR model. It can be seen from the smaller standard error value (0.000398), the larger coefficient of determination (0.5468), and the smaller AIC (-152.52) than the model with the fixed Gaussian kernel weighting.

In this paper, we investigated the evolution of user behavior on Ethereum through building directed and weighted external owned accounts trading networks (DWETN) from August 10th, 2015 to June 9th, 2017. It is showed that the evolution of the structural properties of the DWETN and important events are closely related. The relationship between the number of users and the total number of transactions is linear, and each user will have 1.5 transactions on average. Out-degree, in-degree, out-strength, and in-strength distributions follow the power-law distributions, and the power exponent values of out-degree and in-degree distributions change much more than out-strength and in-strength distributions, which implies that, despite the changes in network structure, the transaction behavior pattern of users have inherent stability. The evolution of directed degree correlation coefficients and directed and weighted Pearson correlation coefficients of the network indicates that in-degree(in-strength) of the source node of an edge has a weak effect on the target node's in-degree(in-strength) and out-degree(out-strength), and as time evolves, the effect of out-degree(out-strength) of the source node of an edge on the target node's in-degree(in-strength) and out-degree(out-strength) from negative to weak.

The ability to predict crop yield is vital for food security worldwide and forecasting crop prices can help farmers avoid price crash. In this work, an investigation of using satellite images and deep learning models to predict crop yields as well as forecasting farmers' prices is conducted. For tractability, dimensionality reduction is achieved by converting the images to histograms representing the pixel frequency. The models tested are LSTM, CNN, CNN-LSTM, CNN-LSTM ensemble as well as a Gaussian Process added to each for enhanced performance. It is found that the proposed ensemble of CNN-LSTMs is the best in predicting the yearly soybean yields in addition to forecasting the daily strawberry yields and prices. It outperforms models suggested in the literature with an improvement of 31% in terms of average Root Mean Square Error (RMSE).

Monitoring crop phenology significantly assists agricultural managing practices and plays an important role in crop yield predictions. Multi-temporal satellite-based observations allow analyzing vegetation seasonal dynamics over large areas by using vegetation indices or deriving biophysical variables. This study presents a framework for automatic corn phenology characterization based on high spatial and temporal resolution time series. By using the Difference Vegetation Index (DVI) estimated from Sentinel-2 data over Iowa (US), independent phenological models were optimized using Gaussian Processes regression. Their respective performances were assessed based on simulated phenological indicators estimated with double logistic approach. The results showed good model performances for the estimation of different phenological phases such as Start-of-Season and End-of-Season, with a mean RMSE and $R^{2}$ values of about 3.50 days and 0.86, respectively, and this with a gain in runtime of about 380 times faster than the double logistic method. To the benefit of crop monitoring community, all these findings will be implemented into the freely downloadable GUI toolbox DATimeS (Decomposition and Analysis of Time Series Software - https://artmotoolbox.com/).

Predicting biophysical parameters with high accuracy and fast speed based on remote sensing-based modeling is an attractive topic. In this context, the revisit time, coverage, and illumination condition in-dependency make the Polarimetric Synthetic Aperture Radar (PoISAR) data is an attractive tool. In this paper, one of the most studied biophysical parameters, Leaf Area Index (LAI), is chosen to assess Polynomial Chaos Expansion (PCE) regression, commonly used metamodeling due to its precise and rapid approximation performance. Experimental analysis based on AgriSAR 2009 campaign, including oat and canola, is given to validate the PCE in the regression. According to the accuracy analysis, the Pearson correlation of 88% and 95% for oat and canola, respectively, were achieved.

In recent years, the breast cancer classification has piqued attention in the area of health-care informatics which is the second leading cause of mortality in women due to cancer. The spread of Breast Cancer growth and its high casualty has prodded a ton of research for contemplating its causes and medicines. The enormous number of genes and their entwining relations requires progressed AI models, instead of simple statistical and correlation analysis. Having the objective to propel the present status of knowledge concerning the early determination of breast cancer, we utilized Random Forest (RF), Logistic Regression(LR), Support Vector Machine(SVM), and Gaussian Process(GP) classifiers, joined with testing unique and novel biomarkers. This paper presents a correlation of these Machine Learning (ML) algorithms by estimating their order test exactness, and their sensitivity and specificity esteem. For the execution of the ML calculations, the dataset was divided into 80% for the training phase, and 20% for the testing phase. Results show that Gaussian Process performed well with 90% test exactness on the order task.

Objective To study the transmission parameters of hand-foot-mouth (HFM) disease and the effects of climate on the incidence, by focusing on the HFM infection in Guangxi Province. Method A TSIR model was employed to analyze the transmission parameters. Pearson correlation, multiple linear regression model, generalized additive model and random forest model used to reveal the effects of climate on HFM disease transmission. Results The result indicates that the reporting rate and the infection rate were 0.37 and 0.00035, respectively, with the size of susceptible individuals as about 490 thousand. Since Guangxi has a subtropical monsoon climate, the study shows that the strongest influence on HFM infection is temperature and followed by rainfall through the Pearson correlation coefficients as 0.55 and 0.4, respectively. Incidence linearly grows with increase of temperature, but the rainfall-HFMD relationship was shown as an approximately inverted U shape curve, with a threshold as 10.5mm. Conclusion Low reporting rate, high infection rate, and huge susceptible people, indicated the severity of HFM infections in Guangxi Province. Evident influence of climate on the incidence illustrates that climate can be used as an indicator for the disease control.

Rice is a staple cuisine for people in Asia, particularly in Indonesia. Rice reserves that are optimal provide sufficient dietary needs. The fundamental issue is that rice quality has deteriorated in recent years, resulting in losses. The traditional method for determining rice quality is to utilize human senses to detect the aroma of rice and examine the texture of the rice. In this paper, we use electronic nose dataser to predict rice quality using the Naive Bayes classifier methods; Gaussian, Bernoulli, Multinomial, and Complement. Among these four methods, Multinomial obtains an accuracy of 97%, Complement obtains an accuracy of 98%, and Gaussian obtains an accuracy of 82%, these methods showed a high ROC value of 1.00, indicating perfect classification. With a ROC value of 0.54, the Bernoulli method exhibited the least performance in classifying the rice quality by obtaining an accuracy of 52%• The managerial implications of these findings are discussed.

This paper investigates the impacts of increasing levels of electric vehicle (EV) penetration in an urban distribution network of Surabaya, Indonesia. Surabaya is the second most populous city in Indonesia where the network is operated by national electric company of Indonesia (PT PLN Disjatim) The investigation is conducted under concentrated and distributed EV penetration (each with 10% and 20% penetration calculated as a percentage of residential load). Loads are modeled as dependent variable of the hour selection within the day and added with a normally distributed uncertainty on each of selected load value to better replicate uncertainty within the daily loading. Whereas, EV charging pattern is modelled as probabilistic uncertain sources using Markov Chain Monte Carlo to better replicate the practical condition. Voltage at each node and aggregate losses are employed to quantify and contextualize the impact. Maximum EV penetration where it violates the statutory threshold is also explored in this paper. Simulation result shows that Surabaya distribution network able to withstand EV penetration in both scenarios with the maximum EV penetration is almost doubled than the highest penetration level from simulated cases.

Power system operations are becoming more challenging with the increasing penetration of renewable-based re- sources such as photovoltaic (PV) generation. In this regard, obtaining accurate solar power output forecasts allows a deepening penetration of renewable-based resources in a secure and reliable way. In this paper, we propose a probabilistic framework to predict short-term PV output taking into account the uncertainty of weather data as well as the variability of PV output over time. To this end, we use datasets comprising of meteorological weather data such as temperature, irradiance, zenith, and azimuth and solar power output. We cluster these data in categories and train a Matern 5/2 Gaussian Process Regression model for each cluster. More specifically, we cluster the data into one to eight different partitions by making use of the k-means algorithm. In order to identify the optimal number of clusters we use the Elbow and Gap methods. We compare the results obtained for the different number of clusters with the (i) 5-fold cross-validation; and (ii) holding out 30 representative days as test data. The results showed that the optimal number of clusters is four, since in comparison to higher number of clusters the increase in the forecast error was marginal.

In this paper, our goal is to leverage cellular network traffic data to model and forecast the number of COVID-19 infections in the future. To this end, we partner with one of the main cellular network providers in Brazil, TIM Brazil, and collect and analyze cellular network connections from 973 antennas for all users in the city of Rio de Janeiro and its suburbs. We develop a Markovian model that captures the mobility of individuals across municipalities of the city. The transition probabilities of the Markov chain are determined by analyzing user-level mobility events between antennas from the cellular network connectivity logs. We combine the aggregate mobility characteristics across municipalities as evidenced from the transition probabilities with the number of reported COVID-19 cases in a municipality during a particular week to design mobility-aware COVID-19 case prediction models that predict the number of cases for the following week. Our experiments demonstrate that our mobility-aware models significantly outperform a baseline mobility-agnostic linear regression model in terms of metrics such as Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE).

Non-Technical Losses (NTL) represent losses due to electricity theft and meter malfunction and are a big problem for electrical grid and electrical sector agents. Previous studies on the subject show that electricity consumption is directly associated with temperature variations, and the non-linearity of these curves can be linked to illicit activities. Thus, the objective of this work is to perform a descriptive analysis related on consumption data of a Brazilian utility, associated with exogenous data (ambient temperature), evaluating the importance of this association on detecting and identifying possible fraudulent consumers. To be assertive, three different consumer units were identified from the inspection history, which were classified as consumer without NTL, consumer with NTL, and a suspected case. Statistical analysis of datasets was perform using Pearson Correlation Technique. In addition, a graphical analysis of possible irregularities was performed.

The correlation analysis of weather factors and outage duration can provide reference for enterprises to make decisions in advance when weather changes, which guarantees stable operation of power grid system. Therefore, aiming at the correlation between weather factors and outage duration, this paper researches linear correlation and nonlinear correlation between them from Pearson correlation coefficient and distance correlation coefficient. Secondly, the gbdt algorithm model is used to measure the contribution ratio of different weather factors in the prediction model by establishing the prediction model between weather influencing factors and response variables, so as to obtain the correlation between different weather factors and outage duration. Finally, the experiment is carried out on a real data set. According to the analysis of the experimental results, the correlation between weather factors and outage duration is described.

Human behavior plays an important role in social, cyber-physical, socio-cyber-physical systems. Therefore the estimates of a person's behavior parameters are needed in the area of risk management of such systems. The gamma Poisson model of behavior is used when only limited data on the several last episodes is available, e.g. when the data is collected via self-reports. The paper presents the Cox regression for the problem of the intensity parameters estimation under gamma Poisson model of behavior. The model is illustrated with the data on public posting in online social media.

In order to improve fire recognition, a forest fire feature extraction algorithm based on color feature statistics is proposed, which combines manual segmentation and statistics of the mean and variance of each channel in RGB color space to construct a Gaussian distribution probability density function. Set a reasonable threshold according to the probability density distribution of the fire area image, and then the threshold is used to judge the probability density of the sample image. Experimental results show that this method can quickly and effectively identify the flames of forest fires in complex scenarios.

In this paper, a process yield-index for autocorrelation between nonlinear profiles is proposed. Applying a one-dimensional Taylor series expansion, the mean and variance of the estimator of the yield index are derived. To evaluate the performance of the proposed yield index a simulation study is conducted. The new index is employed to design three acceptance sampling plans for quality characteristics described by auto-correlated nonlinear profiles. The first sampling plan is based on resubmission. The remaining two follow the repetitive group and multiple dependent repetitive sampling schemes respectively. For all sampling plans, the operating characteristic function is developed. A non-linear optimization approach with search algorithm is employed to determine the number of profiles required for inspection and to decide critical values for acceptance, rejection, and resubmission criteria. The performance of the new methods is investigated and compared with traditional single sampling plan. The comparison confirms all of the three proposed methods possess higher efficiency than a single sampling plan in terms of sample size. For practical applications tables are provided. Two numerical examples from the automobile engine testing and particleboard manufacturing process are used to demonstrate the applicability of the proposed methods.

Mid-term load forecasting (MTLF) is crucial for power transmission system planning, safe operation, and maintenance. It's very significant in a developing country like Cambodia. In the past decade, many kinds of research have been developed for MLTF. However, the specific research of MLTF for Cambodia's transmission system is still insufficient to meet the optimal planning for the rapid growth of the demand. This is the reason that Cambodia would require more research work to be contributed to this field. This paper presents an attempt at a medium-term forecast (one year ahead prediction) for the transmission system load of the Electricite Du Cambodge (EDC) company in Cambodia with the monthly index of historical data from 2010 to 2019, using a hybrid model of combining the Seasonal ARIMA (SARIMA) and Gaussian Process Regression (GPR) model. The data sets were split into two sets of train (2010–2018) and test (2019) data. The proposed model procedure had two main steps which the first step was to train the SARIMA from train data sets then simulated the residual values and estimated the load values. Next, we summed the residual values to the system load values in the train data sets and used that summed up values as new variables to train the kernel-based GPR model while the estimated load from SARIMA was kept as another new variable for simulating the trained GPR model. The other variables such as dates, months of the years, relative humidity, wet-bulb temperature, and total non-working days of each month also were taken into account to train with the GPR model. Three different kernels function were tested in this research work and the modelings were conducted on Matlab 2019b. The Mean Absolute Error (MAE) and the Mean Absolute Percentage Error (MAPE) is used as the indicator to evaluate the performance of the simulated data from models to the actual data. The obtained results have shown that the MAPE of SARIMA and GPR were 3.46 percent while the proposed model performed by 1.95 percent.

Maximum likelihood estimates and corresponding confidence regions of the estimates are commonly used in statistical inference. In practice, people usually construct approximate confidence regions with the Fisher information at given sample data based on the asymptotic normal distribution of the MLE (maximum likelihood estimator). Two common Fisher information matrices (FIMs, for multivariate parameters) are the observed FIM (the Hessian matrix of negative log-likelihood function) and the expected FIM (the expectation of the observed FIM). In this article, we prove that under certain conditions and with MSE (mean-squared error) criterion, approximate confidence interval of each element of the MLE with the expected FIM is at least as accurate as that with the observed FIM.

2021 International Conference on Electric Power Engineering - Palestine, ICEPE-P 2021 -- 23 March 2021 through 24 March 2021 -- Many distribution functions for representing the wind power potential have been proposed. The fitness of the results mainly depends on the used estimation method and the wind pattern of the analyzed area. The selection of a convenient statistical distribution for characterizing wind speed distribution is a critical factor. This paper utilizes three well-known statistical distributions, namely, Weibull, Poisson, and Lognormal to model the wind power in Catalca in the Marmara area located in Turkey. The parameters of these distributions are optimized based on the Genetic Algorithms optimization. The real data of Catalca which was obtained from the national metrology station for three years, are statistically analyzed at 30, 60, and 80 m heights. Root mean square error, correlation coefficient, and mean absolute error measures are exploited to show distributions accuracy differences. Based on the obtained results, the Weibull distribution is superior to others in modelling the real data of Catalca in terms of all used accuracy measures. © 2021 IEEE.

An accurate and careful analysis of posture changes for a dairy cow prior to calving plays an important role in making calving time prediction. The patterns of activities such as frequent changes in postures of a pregnant cows during the time closer to calving are utilized as indicators to predict the time of calving. In this paper, we introduce Markov Chain Monte Carlo (MCMC) method to generate the patterns of four states activities such lying, transitions from lying to standing, standing itself and transitions from standing to lying based on the monitored cow activity changes data three days prior to calving. The validity of the generated cow activities in posture changes data is compared with the actual collected data in terms of Euclidean and Cosine distance measures. The experimental results show that the method in this paper can be used as a generalized method to generate synthetic data series of dairy cow activities prior to calving.

Given the difficulty of accurate prediction for the battery’s state of health (SOH), an integrated method based on ensemble learning for SOH prediction is proposed in this paper. First, the energy of the charging fragment is selected to express the battery SOH. Second, the feature of the charging fragment is extracted by using the Pearson correlation coefficient. Third, the four machine learning algorithms including multiple linear regression, support vector machine, random forest, and random sample consensus regression are combined through ensemble learning to evaluate the SOH. Symmetric mean absolute percentage error (SMAPE), mean absolute error (MAE), mean absolute relative error (MARE), and mean square error (MSE) is selected as the evaluation index. Especially, the output of the proposed algorithm and the other four machine learning algorithms are compared. The results show that the proposed algorithm has the lowest error, which can predict the SOH well and have better stability.

In this paper, we propose an approach to construct uniform error bounds (or confidence intervals) for stochastic kriging with a prescribed confidence level. The theoretical development sheds some light on the impact of simulation experimental designs and budget allocation schemes as well as their relative importance on the large-sample properties of stochastic kriging. Through numerical evaluations, we demonstrate the superiority of the uniform error bounds to the simultaneous confidence intervals obtained by applying Bonferroni correction under various experimental settings.

Multivariate Poisson (MP) counts are common in the course of manufacturing and service process. It is significant to monitor the MP counts and judge whether the process is in control or not. Most of the previous researches assumed that the variables of each univariate Poisson process are independent. Taking the temporal and spatial correlations into account, this article proposes an integrated model based on copula model and autoregressive (AR) process. Furthermore, the inference functions for margins (IFM) method and the expectation maximization (EM) algorithm accompanied by sequential importance resampling (SIR) method, provide satisfactory estimators in the proposed model.

This paper proposes a formula based on Pearson Correlation Coefficient for the calculation of weights of attributes. These weights are used in a version of k-Nearest Neighbor (k-NN) method called Attribute Weighted k-NN with the aim to achieve better performance than simple k-NN and Distance Weighted k-NN by eliminating the dominating effect of irrelevant attributes while calculating the distance between instances. The proposed formula of weights computation in Attribute Weighted k-NN also performs better than when weights are assigned the same as Pearson Correlation Coefficient values. Weights computed using the proposed formula decrease the Root Mean Square Error (RMSE) by 9.26%, 2.63% and 2.51% and increase the R-Squared by 1.36%, 0.25% and 0.2% for NO, NO 2 and NOX, respectively, in comparison to the weights assigned the same as Pearson Correlation Coefficient values in Attribute Weighted k-NN. This paper has also been able to reduce the execution time of Attribute Weighted k-NN and make it near to that of simple k-NN and Distance Weighted k-NN. Before applying k-NN, data preprocessing steps like Data Cleaning (Missing Values and Noisy Data Removal) and Attribute Reduction are applied to the data to transform it into an efficient and useful format.

Accurate and efficient power demand forecasting in urban settings is essential for making decisions related to planning, managing and operations in electricity supply. This task, however, is complicated due to many sources of uncertainty such as due to the variation in weather conditions and household or other needs that influence the inherent stochastic and nonlinear characteristics of electricity demand. Due to the modeling flexibility and computational efficiency afforded by it, a Gaussian process model is employed in this study for energy demand prediction as a function of temperature. A Gaussian process model is a Bayesian non-parametric regression method that models data using a joint Gaussian distribution with mean and covariance functions. The selected mean function is modeled as a polynomial function of temperature, whereas the covariance function is appropriately selected to reflect the actual data patterns. We employ real data sets of daily temperature and electricity demand from Austin, Texas, USA to assess the effectiveness of the proposed method for load forecasting. The accuracy of the model prediction is evaluated using metrics such as mean absolute error (MAE), root mean squared error (RMSE), mean absolute percentage error (MAPE) and 95% confidence interval (95% CI). A numerical study undertaken demonstrates that the proposed method has promise for energy demand prediction.

On 30 July 2020, a total number of 301,530 diagnosed COVID-19 cases were reported in Iran, with 261,200 recovered and 16,569 dead. The COVID-19 pandemic started with 2 patients in Qom city in Iran on 20 February 2020. Accurate prediction of the end of the COVID-19 pandemic and the total number of populations affected is challenging. In this study, several widely used models, including Richards, Gompertz, Logistic, Ratkowsky, and SIRD models, are used to project dynamics of the COVID-19 pandemic in the future of Iran by fitting the present and the past clinical data. Iran is the only country facing a second wave of COVID-19 infections, which makes its data difficult to analyze. The present study's main contribution is to forecast the near-future of COVID-19 trends to allow non-pharmacological interventions (NPI) by public health authorities and/or government policymakers. We have divided the COVID-19 pandemic in Iran into two waves, Wave I, from February 20, 2020 to May 4, 2020, and Wave II from May 5, 2020, to the present. Two statistical methods, i.e., Pearson correlation coefficient (R) and the coefficient of determination (R2), are used to assess the accuracy of studied models. Results for Wave I Logistic, Ratkowsky, and SIRD models have correctly fitted COVID-19 data in Iran. SIRD model has fitted the first peak of infection very closely on April 6, 2020, with 34,447 cases (The actual peak day was April 7, 2020, with 30,387 active infected patients) with the re-production number R0=3.95. Results of Wave II indicate that the SIRD model has precisely fitted with the second peak of infection, which was on June 20, 2020, with 19,088 active infected cases compared with the actual peak day on June 21,2020, with 17,644 cases. In Wave II, the re-production number R0=1.45 is reduced, indicating a lower transmission rate. We aimed to provide even a rough project future trends of COVID-19 in Iran for NPI decisions. Between 180,000 to 250,000 infected cases and a death toll of between 6,000 to 65,000 cases are expected in Wave II of COVID-19 in Iran. There is currently no analytical method to project more waves of COVID-19 beyond Wave II.

Eruption of COVID-19 patients in 215 countries worldwide have urged for robust predictive methods that can detect as early as possible size and duration of the contagious disease and also providing precision predictions. In many recent literatures reported on COVID-19, one or more essential parts of such investigation were missed. One of crucial elements for any predictive method is that such methods should fit simultaneously as many data as possible; these data could be total infected cases, daily hospitalized cases, cumulative recovered cases and deceased cases and so on. Other crucial elements include sensitivity and precision of such predictive methods on amount of data as the contagious disease evolved day by day. To show importance of these aspects, we have evaluated the standard SIRD model and a newly introduced Gaussian-SIRD model on development of COVID-19 in Kuwait. It is observed that SIRD model quickly pick up main trends of COVID-19 development; but Gaussian-SIRD model provides precise prediction at longer period of time.

Humanity's quest to comprehend nature has led to increasing demand in predictive sciences. Much of this is due to the exponential growth of energy production and the ever increasing power prices. Regions with power productions that are carried out with imported coal and oil will be more self-sustaining if alternative sources of power generation such as wind power are discharged. In this paper two predictive models, Markov Chain and Regression Analysis are implemented to predict wind power. In the first model, Markov Chain with 15 and 30 states has been constructed for the short-term forecast of wind power. The second method is a second order polynomial regression with independent variable as wind speed and dependent variable as wind power has been implemented for the medium-term forecast of wind power. A comparative study of both the models has been done to know the best model that suits the forecast. The geographical region under study is a wind farm in Chitradurga, Karnataka. The wind speed values have been sampled at an interval of 10 minutes for a period of three years. Various forecasting errors have been enumerated to audit the credibility of the models.

The most dedicated mathematical tool for spatial traffic modeling is based on stochastic geometry and consider base stations (BSs) and the user equipment (UE) locations modeled as two independent Poisson point processes (PPP). This approach is a relatively good model for generating BS locations but has a major drawback in modelling UE locations due to the lack of adjustability in representing the severity of nonuniformity of the UE position. The second issue is represented by the lack of correlation between the two independent PPPs. The present work reviews different adjustable traffic statistics used in heterogenous traffic generation, along with a novel proposed statistic: the spatial autocorrelation of the users. Only two parameters are used for traffic generation (Coefficient of Variation and the Correlation Coefficient between UE and BS). The proposed traffic generation model capabilities are illustrated in a single tier wireless cellular network.

We propose a method to determine photon-number statistics of excess noise in laser radiation from measured photocounting statistics. The method based on the multi-objective optimization approach is applied to blind deconvolution problem to determine excess noise distribution from the convolution of this one and Poisson photon-number distribution of laser radiation.

We have previously shown how the Bayesian Fisher Information (BFI) is related to the average of the Minimum Probability of Error (MPE) for the task of detecting a small change in a parameter. The prior probabilities for the two hypotheses corresponding to two nearby values of the parameter are derived from the prior distribution on the parameter, as they are for the Ziv-Zakai inequality. The average in this result and others to be discussed is performed over the parameter values using the prior. We extend this result to the case where a cost function is assigned to the detection task. In this case we determine the lowest order approximation of the average of the Bayesian Risk for the detection task. We have shown that there is a similar relation between the average Shannon Information (SI) for this detection task and the BFI, and that there is an integral relation between SI for any detection task and the MPE for that task. We will show the Bayesian perspective on this last result, and how these two results about SI are related.

Climate change, as an important environmental issue, has been widely investigated in recent decades. On the one hand, the climate prediction is an essential part for policy makers to response to the change of climate, which has received many attentions. On the other hand, there is another challenging problem facing us today that some abnormal weathers occur globally, which seems to have relation to climate change, e.g., the global greenhouse effect, but with little existing researches on this relation. Therefore, in this paper, we propose two kinds of climatic and meteorological models based on statistical data: 1) an autoregressive-moving-average (ARMA) prediction model with principal component analysis (PCA) and 2) abnormal analysis model based on Pearson correlation coefficient (PCC). In detail, firstly, we propose the PCA-ARMA prediction model to predict climate change in the next 25 years, including two steps: 1) generation of new components for data reduction by PCA using the past 75 years' data, and 2) prediction based on step 1 by ARMA for next 25 years. Then, we establish another model to find out the relation between climate change and abnormal weathers, e.g., the extreme cold weather, mainly by PCC. The relevant data are collected, and by these two models, we get the corresponding results, which show that our prediction fits well and the abnormal weather is strongly connected with the climate change.

This article describes analytical work carried out in a pilot project for the Swedish Space Data Lab (SSDL), which focused on monitoring drought in the Malardalen region in central Sweden. Normalized Difference Vegetation Index (NDVI) and the Moisture Stress Index (MSI) – commonly used to analyse drought – are estimated from Sentinel 2 satellite data and averaged over a selection of seven grassland areas of interest. To derive a complete time-series over a season that interpolates over days with missing data, we use Gaussian Process Regression, a technique from multivariate Bayesian analysis. The analysis show significant differences at 95% confidence for five out of seven areas when comparing the peak drought period in the dry year 2018 compared to the corresponding period in 2019. A cross-validation analysis indicates that the model parameter estimates are robust for temporal covariance structure (while inconclusive for the spatial dimensions). There were no signs of over-fitting when comparing in-sample and out-of-sample RMSE.

Indonesia is a country with a very large population and a fairly high population growth. Population growth is one of the important indicators in demography that will affect the availability of food, land and employment. Hence, population growth data is very important to government when designing development policies. The fuzzy time series models have been widely used in previous studies to forecast enrollment data, stock exchanges and others. In this study, three types of fuzzy time series models were used namely Chen, Cheng, and Markov Chain models to predict the total population in Tanjugpinang city, Riau Islands Province. The models are compared based on error analysis, namely: Mean Average Percentage Error (MAPE) and Root Mean Square Error (RMSE) values. The results showed that Markov chain model yielded a very low MAPE, namely 0.0457%. The MAPE value for the Cheng model is 0.0535%, while the MAPE value for the Chen model as the reference model is 0.1428%. Based on the RMSE calculation, the Markov chain gives the best results with a value of 145, RMSE for Cheng 149 and the Chen model gives an RMSE value of 345. The best model obtained for population forecasting is the Markov chain as it has the smallest RMSE, and also the Markov chain model is the most accurate model in the used dataset based on the MAPE value.

In this study, made from observations in 4 different locations in Doha, the first location is close to the education campus with an open and flat area, the second location is the trade zone, the third location is a region with high buildings near the city center, and the fourth location is selected as the factory and workshop area. Three different algorithms, Bayes Based Analysis, Largest Sense Estimation, and Naive Bayes classifier, were used to estimate and predict spectrum usage on a data set, with 1 minute observation for each location, with a total of 4320 different readings in the 700 - 3000 MHz spectrum. The optimum solution was sought for the prediction and estimation of its use. After the optimum method was chosen, the relationship of the chosen method with the past and previous situations was examined. As a result of these algorithms, Bayes Based algorithm was chosen as the most suitable algorithm and performance was measured as %88:94.

A one-dimensional discrete random variable X describes two populations. A random sample is drawn from each. We formalize a generic statistical test to determine whether the samples provide enough evidence that the distributions of X in the two populations are different. After showing the drawbacks of the existing analytical tests, we develop a Bootstrap procedure with Pearson test statistic calculated from a contingency table. The p-value is estimated using the simulated conditional distribution of the test statistics under null hypothesis for equality of population distributions. Compared to analytical tests, our procedure gives higher precision and decreases uncertainty in both small and big samples. We apply our procedure to survey data on leadership capabilities in academia among two samples of respondents. We take one question from the survey and analyse significance of differences in the responses depending on three criteria.

Correlation power analysis (CPA) is a side- channel attack (SCA) which exploits the information leaked through the power supply current and voltage, or the electromagnetic emissions of the attacked digital system. It uses statistical analysis of a large number of power supply measurements to retrieve the secrets of the digital system. Correlation power analysis uses a number of hypothetical secret keys which are correlated to the measurements of the attacked system. Usually correlation power analysis uses the Pearson correlation coefficient, but the intermediary values and the power supply measurements can have a nonlinear relationship. The paper investigates the application of the distance correlation in the correlation power analysis and compares it to the Pearson correlation coefficient. The comparison is based on a side-channel attack on a multiplication operation of an input message and a secret key. The results of the comparison show that the distance correlation achieves a higher prominence of the correct secret key than the Pearson correlation coefficient.